Frank Werblin

Research Interests

Discovering the language of vision: How is the visual world represented in the brain? What mechanisms mediate the generation of these representations? What representations would be needed to provide a rich visual environment for the blind?

Research Summary

Computational and Physiological Studies of Image Processing in Retina and Silicon

We recently discovered a set of more than a dozen different representations that are formed in parallel in the mammalian retina. This set constitutes a visual language initiated in the retina and carried to higher visual centers for further processing elaboration and refinement. Our goal is to understand this fundamental visual language, decipher it grammar understand its meaning, determine how it is formed. Our approach is to uncover its secrets with physiological approaches, perform a variety of "what if" experiments through simulating its behavior, determine how it operates on natural scenes, and build a "visual camera" that simulates visual function and can be used as the front end of a visual prosthetic system.

Selected Publications

Werblin, F., Roska, B. & Balya, D. (2001) Parallel processing in the mammalian retina: lateral and vertical interactions across stacked representations. In Kolb, Ripps and Wu (Eds.) Progress in Brain Research 131: 229-238.

Balya. D., Roska, B., Roska, T. and Werblin, F.S. (2001) A CNN Model of a mammalian Retina. Int. N. Circuit Theory and Applications (in press)

Roska, B. & Werblin, F.S. (2001) Vertical Interactions Across Ten Parallel Stacked Representations in the Mammalian Retina Nature 410: 583-587.